Jun 26, 2016 - Division of Earth and Ocean Sciences, Nicholas School of the Environment, Duke University, Durham, USA. Received ... died scale in urban climate research. ..... (KRDU) that has nearly 400 flights per day and three runways.
Journal of Environmental Protection, 2016, 7, 1072-1088 Published Online June 2016 in SciRes. http://www.scirp.org/journal/jep http://dx.doi.org/10.4236/jep.2016.77096
Climate & Sustainability Implications of Land Use Alterations in an Urbanizing Region: Raleigh-Durham, North Carolina E. M. B. Doran, J. S. Golden Division of Earth and Ocean Sciences, Nicholas School of the Environment, Duke University, Durham, USA Received 18 April 2016; accepted 26 June 2016; published 29 June 2016 Copyright © 2016 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/
Abstract Urban climate is the most immediate manifestation of the warming global climate for the majority of people on earth. Nearly half of those people live in small to medium sized cities, an understudied scale in urban climate research. Widespread characterization would be useful to decision makers in planning and design for land use decisions. Using a multi-method approach, the mesoscale UHI in the study region is characterized and the secular trend over the last sixty years evaluated. Under isolated ideal conditions the findings indicate a UHI of 5.3˚C ± 0.97˚C to be present in the study area, the magnitude of which is growing over time.
Keywords Land Use, Urban Heat Island, Mesoscale, Urban Climate, Piedmont, North Carolina
1. Introduction Global climate is an area of active research and ongoing attention by the international scientific community [1] [2]. With more than half the world’s current population living in cities as of 2007 [3], a trend that is projected to continue to increase, urban climate is the most immediate manifestation of the changing Earth for the majority of people on the planet [4]. In areas where urbanization and significant land use change is occurring, evidence suggests the strength of urban climate warming signals is growing faster than the global mean climate [5] [6] which can have significant environmental implications such as increased water and energy consumption [7] and on air quality [8]. Since the modern era of scientific observations and Luke Howard’s early 1800s work in London [9], the Urban Heat Island (UHI) has been observed in cities of various sizes around the world [10] [11]. The UHI is deHow to cite this paper: Doran, E.M.B. and Golden, J.S. (2016) Climate & Sustainability Implications of Land Use Alterations in an Urbanizing Region: Raleigh-Durham, North Carolina. Journal of Environmental Protection, 7, 1072-1088. http://dx.doi.org/10.4236/jep.2016.77096
E. M. B. Doran, J. S. Golden
fined as the warmer nature of the urban environment compared to the surrounding rural environs, including surfaces and canopy and boundary layer air masses [9] [10]—a manifestation of land use and land cover decisions and actions. At the same time anthropogenic emissions are driving the rise in global mean temperatures, the overall growth and rapid urbanization of the planet’s population—from 30 percent urban in 1950 to 54 percent urban in 2014 while nearly tripling the total population—are driving a transition from native vegetation and land cover types to engineered infrastructure and land typologies that in some cases have dramatically different thermodynamic properties [4] [12]-[14]. In the United States, UHI studies have largely focused on characteristically dense urban cities where UHI signals are the strongest. In the arid southwest, for example, researchers have used the city of Phoenix, AZ to investigate various aspects of the physical UHI phenomenon including the correlation of magnitude with population, as well as impacts and mitigation opportunities [15]. UHI research has also centered around the humid sub-tropical cities of Houston, TX [16] [17] and Atlanta, GA [18]. The city of Houston, TX, as the fourth largest city in the country by both population and land area [19], has drawn interest from researchers demonstrating specific remote sensing [17] and regional scale climate modeling [16] methods, while the Atlanta, GA UHI has been found to exhibit strong perturbation effects on the local hydrologic cycle with increased storm events concentrated in and affecting specific parts of the city [18]. Similarly, in the coastal city of New York, NY the urban heat island has some potential to stimulate the formation of a sea breeze front due to differential heating between the urban environment and adjacent ocean [20]. Studies linking the impacts of urban heat islands to socioeconomic conditions have also focused on larger cities including Philadelphia, PA, finding that social and economic indicators can be used to identify geographic areas with higher risk to morbidity and mortality [21]. The studies referenced here are not intended to be exhaustive of all UHI research but be representative of the majority of the UHI research. Relatively little research has focused on urban areas (smaller than 500,000 residents) and small cities (500,000 to 1 million residents) and in regional configurations of multiple small connected cities, despite the fact that cities with populations under 500,000 were home to almost half the world’s urban population as of 2014 [3]. While fully two thirds of Europeans live in these small urban areas, the largest growth is occurring in Africa and Asia where more than half of the population lives in small or medium urban areas of less than one million residents [3]. Although understudied in the most common context, there is increasingly clear evidence of the consequences of land use alterations. A recent review summarizing 50 years of UHI research in 225 cities, suggests the UHI results in increased consumption of energy, higher levels of urban environmental pollution, lower thermal comfort for residents, and increased mortality [22]. Additional reviews of recent literature publications concur; UHI impacts include adverse effects on human health including morbidity and mortality [23] [24]; environmental degradation of adjacent natural ecosystems including water bodies, and wildlife; and economic impacts arising from compromised human comfort profiles [6] [11]. To begin to address the nexus of land use decisions and environmental impacts (climatic) in small and medium sized, regionally connected cities, this study focuses on a rapidly growing region of the United States where the impact of the findings has the potential to support policy-makers around the globe.
2. The Piedmont Region The study region is shown in Figure 1 and comprised of three counties, Wake, Durham and Orange which cover 4000 square kilometers in the central Piedmont region of North Carolina centered around 35.8 latitude, −78.8 longitude. The cities of Raleigh, Durham, and Chapel Hill respectively comprise roughly 17 percent of the land area in the combined counties. Topographically, the study area is characterized by rolling hills with an average elevation of 120 m above sea level. The Appalachian Mountains range runs northeast-southwest 150 km to the northwest of the study region while the Atlantic Ocean is 200 km southeast. The area is anchored by three major universities: Duke University in Durham, University of North Carolina in Chapel Hill, and North Carolina State University in Raleigh. The research triangle is also home to the Research Triangle Park which is the largest research park in the United States with more than 200 companies including large technology and pharmaceutical companies, and major government research laboratories that employ more than 50,000 people, approximately 80% of whom work for multi-national companies [25].
1073
E. M. B. Doran, J. S. Golden
Figure 1. Study area and weather station locations.
The study region also serves as the northern terminus of the Piedmont megaregion which extends down the I85 corridor through Charlotte, North Carolina to Atlanta, Georgia [26]. Defining the geographic scale of a megaregion requires consideration of the integration of economic factors, natural resources, ecosystems, transportation and population that extends beyond the standard political boundaries of metropolitan areas [26]. The Piedmont megaregion is characterized by a combination of fast growing smaller regions and older more established regions. For example, the Piedmont megaregion is home to three of the top 15 fastest growing cities in the United States in 2014 including Raleigh, North Carolina which ranked second; Charlotte, North Carolina which ranked eighth; and, Atlanta, Georgia, a city much older than the other two, which ranked twelfth [27]. This is in contrast to the Northeast megaregion, for example, which extends down the I-95 corridor from Boston, Massachusetts to Washington, DC and is comprised of older, slower growth regions, that tend to have longer history, be well established, and are well connected. One of four megaregions identified in the United States Census South region [26], the study area is further part of the fastest growing Census regions in the country [28]. Comprised of 16 states and the District of Columbia, the South region was home to roughly 37% of the total United States population in 2005.
2.1. Projected Future Growth The current growth rates evident in the study area are projected to continue with significant land use and land cover changes and their associated implications for the region as a whole and for the study area specifically. At the regional level, the United States Census South is projected to continue to grow with intensification in regions already identified as megaregions including the Piedmont. Based on analysis done by Bierwagen, Thomas [28] to adapt the IPCC SRES storylines to the United States [29], the South region could grow to encompass between
1074
E. M. B. Doran, J. S. Golden
38% and 39% of the population by 2030 while growing between 22% and 30% overall. In the Raleigh-Durham-Chapel Hill study area, projections of population growth and distribution compiled as part of the 2040 Long Range Transportation Planning report indicate that while employment is likely to see growth in the exiting transportation and urban core areas, new homes are likely to be split between the urban downtown cores, and the regional fringes, contributing to continued sprawl [30]. Growth in Durham County is projected to be stronger, 1.7 percent per year, than growth within the city limits, 1.5 percent, while the City of Raleigh and Wake County are both expected to exceed 100% growth relative to 2000 Census figures, with annual growth rates around 2.2 percent [31]. These projections are comparable to and even outstrip projected rates of urbanization for African and Asian countries [3].
2.2. Regional Climate Characteristics The Raleigh-Durham-Chapel Hill region of North Carolina, experiences a humid subtropical climate [32]. Based on data collected at the ASOS weather station at the Raleigh-Durham International Airport (KRDU) available from the State Climate Office of North Carolina and shown in Figure 2, average maximum temperatures are around 30˚C in the summer months (June-July-August) and 10˚C in the winter (December-January-February), and average minimum temperatures are around 20˚C in the summer months and 0˚C in the winter. Further, precipitation trends are relatively consistent over the last six decades with an average of 2 - 3 mm of rain per day except during the late summer, early fall (July-August-September) when the average precipitation is not only higher (up to 5 - 6 mm per day), but also more variable. Two regional climatological analyses have been conducted in the study region, both focused on extreme heating. The first, an analysis of synoptic scale summer heat events in the Piedmont between 1951 and 1993 [33] ends in the middle of a period of rapid population growth that extends into the 1990s. The second study focuses
Figure 2. KRDU mean climate characteristics summary for study region including temperature (top) and precipitation (bottom).
1075
E. M. B. Doran, J. S. Golden
on the North Carolina heat wave in 2007 which was an exceptional period in which hundreds of daily maximum and daily high minimum temperature records as well as several all-time temperature records in the state were either tied or broken [34]. The two studies indicate that summer heating conditions in particular are influenced roughly half the time by prevailing southwesterly winds around the Bermuda High or downs loping westerly winds off the Blue Ridge mountains. Both high humidity and low humidity type heating events occur in the study region subject to the same synoptic scale conditions with humidity controlled by regional advection or precipitation in the 30-day window antecedent to a heat event [33]. The mixing depth in the boundary layer was also found to be important and to be controlled by the dry soil conditions leading to midafternoon decreases in the heat index despite continued increases in overall temperature [34].
2.3. Previous UHI Research & Research Questions Limited efforts have been made to characterize the UHI in the study area, however the cities of Raleigh and Durham have appeared in regional, multi-city analyses. Chen and Konrad II [33], for instance, report an average temperature difference between the weather station at the Raleigh-Durham Regional Airport (KRDU) and unspecified cooperative observer sites of 0.8˚C during the timespan of their study (1951-1993) but do not report a secular trend. In another case, Durham was included in an evaluation of the UHI for18 cities in the states of North and South Carolina from 1971-2000 finding evidence that UHIs were present, and that their strength could be correlated to population [35]. Finally, the city of Raleigh, NC was included in a modeled study using data from the NASA MODIS satellite of 65 North American cities that found daytime UHI strength is strongly correlated with mean annual precipitation while nighttime UHI strength is correlated with population [36]. Work specific to the study area does include limited attention to drivers and impacts of the UHI. For instance, urban tree cover has long been considered a mitigation approach for reducing the UHI strength despite a lack of clear guidance for design and implementation [37]. Of two studies focused on tree cover in the City of Raleigh, one suggests the correlation of morphological characteristics (e.g. lot size, age of construction) with the distribution of urban tree coverage [38], while the other indicates a negative correlation of urban tree cover with percent Black residents [39] with implications for social justice. An additional study investigating UHI impacts in the study area included findings that suggest urban warming is a key driver of pest abundance and outbreaks on urban trees [40]. Finally, an analysis of emergency room visits across North Carolina during the heat waves of 2007 and 2008 found that temperature effects were greatest for middle aged men, and more commonly residents of rural counties [41]. This finding may be logical given the agricultural nature of the state’s economy, the fact that daytime UHI is often negative, with rural temperatures exceeding those in urban areas, however additional work would be needed to assess time lags, the correlation with maximum instead of mean daily temperatures, and the impact of acute heat-wave events. Given the attention to morphological drivers and impacts and the lack of careful regional characterization, we therefore seek to address the following questions: How strong is the UHI in the study region? Is it growing over time, and is this growth explained by trends in population or land use change? Finally, given the humid subtropical climate of the study region, what is the relative importance of controlling for climatological conditions in these findings?
3. Data Sources and Methodology To characterize the current distribution of the regional UHI, we utilize 1) remote sensing techniques. To determine the magnitude of the regional UHI over time, two methods were used: 2) monthly average minimum and maximum UHI, and 3) diurnal hourly UHI including passive climatological control.
3.1. Remote Sensing Method To begin to characterize the spatial distribution of the UHI across the study region, thermal infrared (TIR) data collected by the Landsat-8 OLI-TIRS satellite was used. Band 10 of the TIR aboard Landsat 8 collects light in the wavelength 10.60 - 11.19 μm. Because interest is exclusively in the relative strength of the regional surface-heat profile from a single image, the image data was corrected to At Satellite Brightness Temperature which does not include an atmospheric correction interference [42]. Conversion formulas are provided by the US Geological Survey and the correction factors are provided in the metadata compiled with the scene data [43].
1076
E. M. B. Doran, J. S. Golden
The TIR Band 10 data is first converted to top of atmosphere (TOA) spectral radiance according to Equation (1) [43]. (1) = Lλ M L Qcal + AL where the TOA spectral radiance, Lλ, has units of Watts∙m−2∙s∙rad−1∙μm−1, and where the band-specific multiplicative rescaling factor, ML, equals 3.3420E−04; the addative rescaling factor AL, equals 0.1000; and where, Qcal, is the quantized standard product pixel values. The TOA spectral radiance is then converted to at satellite brightness temperature using Equation (2) [43].
= T
K2 K ln 1 + 1 L λ
− 273.15
(2)
where temperature, T, is corrected to degrees Celsius, and where the thermal correction constants are as follows: K1 = 774.0853 and K2 = 1321.0789. The transect data obtained from the processed image was smoothed using a 15 point moving average. The smoothed data was used to calculate temperature differences between weather station locations. In addition, the 2011 National Land Cover Database for the study area was aggregated into two land cover classifications: urban and rural, with water appearing separately. Urban areas included all developed land classifications (21 - 24) including developed open space as well as low, medium and high density development [44] [45]. The mean temperature was then calculated for each of the two land cover types (urban and rural) using the spatial analyst tool in the ArcGIS environment.
3.2. Mean Monthly Method Data used in this study was collected from five hourly weather stations and two daily weather stations in and around the study region (see Figure 1). The stations used in the analysis including their characteristics (e.g. station type, elevation and date of first record) are summarized in Table 1 while instrument data at each station network is summarized in Table 2. Criteria for station selection included 1) length of climate record, 2) availability of hourly temperature data, and 3) urban or rural location. In the study region, the weather stations with the longest temporal record are those that only record daily observations. The daily stations are part of the National Weather Service Cooperative Observer (COOP stations) program and record observations of minimum and Table 1. Weather station general information summary. Elevation (m ASL)
Lat
Long
LCZ subclassb
Distance from KRDUc (km)
ASOS-standard
133
35.87764
−78.78747
6D
-
8/2/03
ECONET-tower
76
35.59158
−78.45889
CA
43
Louisburg, NC
1/25/03
AWOS -III
112
36.02335
−78.33027
DA
44
Person County Airport
Roxboro, NC
12/14/00
AWOS -III
186
36.28489
−78.98423
DB
49
Duke Forest
Carrboro, NC
3/30/00
RAWS
172
35.97
−79.09
A
29
COOP-TP
91
35.64083
−78.46333
9D
39
COOP-TP
79
36.10278
−78.30389
9D
50
Station IDa
Station name
Near city
First record
KRDU
Raleigh-Durham Airport
Raleigh/ Durham, NC
7/1/48
CLA2
DAQ Clayton Profiler
Clayton, NC
KLHZ
Franklin County Airport
KTDF NDUK
Network
Hourly stations
Daily stations 311820 315123
Clayton Wtp Louisburg
Clayton, NC
11/1/55
January 1, Louisburg, NC 1893
Notes: aStation Locator: http://climate.ncsu.edu/CRONOS/; bLCZ Classification based on Setwart & Oke, 2012; cLat/Long calculator: http://www.nhc.noaa.gov/gccalc.shtml.
1077
E. M. B. Doran, J. S. Golden
Table 2. Weather station network temperature and precipitation sensor technical specifications. Network
ASOS-standard
ECONET-tower
AWOS-III
RAWS
COOP-TP
Sensor
RTDb
Ceramic capacitance
RTDb
Platinum resistance
MMTSd
Range
−62˚C to 54˚C
−52˚C to 60˚C
−128˚C to 54˚C
−50˚C to +50˚C
−48˚C to 52˚C
Accuracy
RMSEc: 1.1˚F - 7.9˚F
±0.2˚C at −40˚C, ±0.3˚C at 20˚C, ±0.4˚C at 40˚C
RMSEc: of ±1.8˚F
±0.6˚C
±0.3˚C
DRRa
60 min
30 min
60 min
60 min
24 hour
Sensor
Heated tipping bucket
Tipping bucket rain gage
Heated tipping bucket
Tipping bucket
Std Rain Gage or Fischer & Porter
Range
0 - 10 in/hr
Infinite in increments of tips
Infinite in increments of tips
0 - 99.99 in
0 - 8 in(SRG), 0 - 20 in (F&P)
Accuracy
±0.02" or 4% hourly total
±0.5 kt
Within 1% if
